Faculty of Community and Public Health 

CLASSICAL EPIDEMIOLOGY

Homework #4
Cigarette Smoking and Lung cancer case Study
Mayson Bali 1185273

Supervised by: Dr Abdullatif Husseini

Question 1: What makes the first study a case-control study?
1- The study compares the smoking habits of lung cancer patients with the smoking
habits of other patients.
2- Tow population comparison.
3- Cases: have the outcome of interest (smoking)
4- Controls: are non-smokers.
5- The observation of two groups of patients admitted lung cancer.

Question 2: What makes the second study a cohort study?

1- Data for the cohort study were obtained from the population of all physicians.
2- They were followed over a specific time.
3- Information about lung cancer came from death certificates and other mortality
data recorded during ensuing years.
4- Observation of causes of death among smoking physicians and compared to causes
of death among non-smoking physicians.

Question 3: Why might hospitals have been chosen as the setting for this study?
1- Hospitals are more convenience
2- The ease of information collection by researchers and easier to find patients.
3- The availability of medical records that would identify all lung cancer cases.
4- Physicians are able to diagnose different cases of disease.

Question 4: What other sources of cases and controls might have been used?
Sources of cases could be from
1- Doctor offices
2- Cancer registers
3- medical records

Sources of controls could be from

1- volunteers
2- geriatrics
3- Participants selected randomly from different healthcare setting.

Question 5: What are the advantages of selecting controls from the same
hospitals as cases?
1- Easier to collect information.
2- Reduce the time required for the study.
3- If both controls and cases took from the same hospital, they could be from the same
region geographically with the same environmental variables.
Question 6: How representative of all persons with lung cancer are hospitalized
patients with lung cancer?

 Hospitalized patients with lung cancer are not completely representative.

 Not all patients with lung cancer are hospitalized and do not receive
medical care.
 Hospitalized lung cancer are mostly newly diagnosed patients,
patients with current lung cancer status or in late stage lung
cancer.

Question 7: How representative of the general population without lung cancer

are hospitalized patients without lung cancer?
 Not completely representative.
 The reasons or cause of hospitalization could be different.
 The general population without lung cancer are much greater than
the sample of people without lung cancer and hospitalized
 The study does not include healthy smoking people (not sick and do
not have lung cancer)

Question 8: How may these representativeness issues affect interpretation of the

study's results?
 Selection bias: The study does not include healthy smoking people (not sick and
do not have lung cancer).
 Limit generalizability: further research needed to confirm the results.
 There are many other confounders of lung cancer that are not associated with
smoking cigarette can affect the interpretations of a study result.

Question 9: From this table, calculate the proportion of cases and controls who
smoked.
Proportion smoked, cases: the number of cases of people when smoke / total
participants =1350/ 15357 = 0.99=99%

Proportion smoked, controls:  the number of people who smoke but do not have
the cases of the disease / the total number of participants

= 1296/1357=0.96= 96%

Question 10: What do you infer from these proportions?

The results are close nearly all patients in in cases and controls are smokers.
Question 11a: Calculate the odds of smoking among the cases.
Odds of exposed = events/non-events =1350/7= 192.9 times more likely to smoke

Question 11b: Calculate the odds of smoking among the controls.

Odds of un-exposed = events/non-events = 1296/61 = 21.25 times more likely to smoke

Question 12: Calculate the ratio of these odds. How does this compare with the
cross-product ratio?
OR = Odds of exposed/ Odds of un-exposed = 192.9/21.25 = 9.08
Cross-product ratio = 1350*61/1296*7 = 9.1 the same results

Question 13: What do you infer from the odds ratio about the relationship
between smoking and lung cancer?
Smoking people are 9.08 times more likely to have lung cancer than non- smokers

Table 2. Most recent amount of cigarettes smoked daily before onset of the
present illness, lung cancer cases and matched controls with other diseases,
Great Britain, 1948-1952.
Question 14: Compute the odds ratio by category of daily cigarette consumption,
comparing each smoking category to non-smokers

Daily number of # Cases # Controls Odds Ratio

cigarettes
0 7 61 referent
1-14 565 706 (565*61)/(706*7)= 6.97
15-24 445 408 (445*61)/(408*7)= 9.50
25+ 340 182 (340*61)/(182*7)= 16.28
All smokers 1,350 1,296 (1,350*61)/( 1,296*7)
=9.07
Total 1,357 1,357

Question 15: Interpret these results.

The increasing in daily number of cigarettes smoked increases the odds of having cancer
Question 16: What are the other possible explanations for the apparent
association?
1- bias with data collection
2- bias in participants respondents rate
3- bias in The region, location, or health care facility where participants
are selected from
4- smokers could have other risk association with lung cancer

Question 17: How might the response rate of 68% affect the study's results?
It is a good response rate and give good result. Having a low response rate could lead
to bias in study results (non- response bias) but doesn’t affect the validity of data
collection. It give a good idea of association between smoking physician and lung
cancer.

Table 3. Number and rate (per 1,000 person-years) of lung cancer deaths
by number of cigarettes smoked per day, Doll and Hill physician cohort
study, Great Britain, 1951-1961.
Daily Death Person- Mortality Rate ratio Rate
number of from lung years at rate /1000 difference /
cigarettes cancer risk Person- 1000
smoked years Person-
years
0 3 42,800 0.07 referent referent
1-14 22 38,600 0.57 8.14 0.5
15-24 54 38,900 1.39 19.79 1.32
25+ 57 25,100 2.27 32.36 2.2
All smokers 133 102,600 1.3 18.57 1.23
Total 136 145,400 0.94

Question 18: Compute lung cancer mortality rates, rate ratios, and rate
differences for each smoking category. What do each of these measures mean?
1- Mortality rate measures the rate of death from lung cancer over the specific
period.
2- Rate ratio is the incidence of cancer in smokers divided by the incidence of
cancer in non-smokers.
3- Rate differences is how different the mortality rate of smokers from non-smokers
mortality rate.

Question 19: What proportion of lung cancer deaths among all smokers can be
attributed to smoking? What is this proportion called?
AR% = (1.3-0.07)/1.3 = 95% attributed risk for population
Lung cancer death attributed to smoking is called case-specific rate or odds
proportion = 1.33/136*1000 = 9.77 =97.7%

Question 20: If no one had smoked, how many deaths from lung cancer would
have been averted?
 - 133 deaths from lung cancer were from smokers
- 133*0.977 = 130 deaths could be avoided

Table 4. Mortality rates (per 1,000 person-years), rate ratios, and excess deaths
from lung cancer and cardiovascular disease by smoking status, Doll and Hill
physician cohort study, Great Britain, 1951-1961.
Mortality rate per 1,000 person-years
Smokers Non- All Rate Excess Attributable
smokers ratio deaths per risk percent
1,000 among
person- smokers
years
Lung cancer 1.30 0.07 0.94 18.5 1.23 95%
Cardiovascula 9.51 7.32 8.87 2.19 1.3 23%
r disease

Question 21: Which cause of death has a stronger association with smoking?
Why?
Lung cancer has a stronger association with smoking as compare to cardiovascular disease
because the rate here is about 19:1 (95% – 23%)

Question 22: Calculate the population attributable risk percent for lung cancer
mortality and for cardiovascular disease mortality. How do they compare? How
do they differ from the attributable risk percent?
PAR% = (Incidence in entire population - Incidence in unexposed) / Incidence in entire
population
PAR% lung cancer = 0.94-0.07/ 0.94 = 0.93*100% =93%
PAR%   cardiovascular disease = 8.87-7.32/8.87*100%= 17.5%
This is an indication that cigarette smoking is highly associated with lung cancer, mortality
rate as compare to cardiovascular disease.

Question 23: How many lung cancer deaths per 1,000 persons per year are
attributable to smoking among the entire population? How many cardiovascular
disease deaths?
Lung cancer death per 1000 persons = 0.93*0.94 = 0.87 deaths per 1,000 person years.
Cardiovascular disease death per 1000 persons = 0.17*8.87 = 1.54 deaths per 1,000 person
years.
Question 24: What do these data imply for the practice of public health and
preventive medicine?
The rate of death from lung cancer decreases as the number of years since quitting smoking
increases. It is good for Public health prevention strategies and programs.

Question 25: Compare the results of the two studies. Comment on the similarities
and differences in the computed measures of association.
- Both of these studies show that cigarette smoking is highly
associated with lung cancer.
- The number in case control study were smaller than in cohort
study
Question 26: What are the advantages and disadvantages of case-control vs.
cohort studies?
Case-control Cohort
Sample size Small Large
Costs Cheap Expensive
Study time Short Long
Rare disease Advantages Disadvantages
Rare exposure Advantages Disadvantages
Multiple exposures disadvantages Advantages
Multiple outcomes Disadvantages Advantages
Progression, spectrum Disadvantages Advantages
of illness
Disease rates Advantages Advantages
Recall bias Advantages Disadvantages
Loss to follow-up Advantages Disadvantages
Selection bias Advantages Disadvantages

Question 27: Which type of study (cohort or case-control) would you have done
first? Why? Why do a second study? Why do the other type of study?
I will do case-control because sample size are small, cheaper, shorter time. Then do cohort
study to find the disease association and to verify the findings with more intensive study.

Question 28: Which of the following criteria for causality are met by the
evidence presented from these two studies?
yes No
Strong association 
Consistency among studies 
Exposure precedes disease 
Dose-response effect 
Biologic plausibility 